Applied Physics Express

LETTERS Related content

- Improved carrier injection and confinement 2 in InGaN light-emitting diodes containing 6 kW/cm UVC laser threshold in optically pumped GaN/AlGaN/GaN triangular barriers Li-Wen Cheng, Jian Ma, Chang-Rui Cao et lasers achieved by controlling point defect al.

- Optoelectronic device physics and formation technology of nitride semiconductors from the UV to the terahertz Theodore D Moustakas and Roberto To cite this article: Ronny Kirste et al 2018 Appl. Phys. Express 11 082101 Paiella

- AlGaN-based laser diodes for the short- wavelength ultraviolet region Harumasa Yoshida, Masakazu Kuwabara, Yoji Yamashita et al. View the article online for updates and enhancements.

This content was downloaded from IP address 152.7.255.201 on 15/03/2019 at 14:34 Applied Physics Express 11, 082101 (2018) https://doi.org/10.7567/APEX.11.082101

6kW/cm2 UVC laser threshold in optically pumped lasers achieved by controlling point defect formation

Ronny Kirste1,2, Qiang Guo1, J. Houston Dycus1, Alexander Franke1, Seiji Mita1,2, Biplab Sarkar1, Pramod Reddy1,2, James M. LeBeau1, Ramon Collazo1, and Zlatko Sitar1,2

1Department of Materials Science and Engineering, North Carolina State University, Raleigh, NC 27695-7919, U.S.A. 2Adroit Materials, 2054 Kildaire Farm Road, Suite 205, Cary, NC 27518, U.S.A.

Received June 22, 2018; accepted July 10, 2018; published online July 27, 2018

Optically pumped lasing from AlGaN/AlN multiple quantum wells grown on single-crystalline AlN substrates with lasing thresholds as low as 6kW/cm2 is demonstrated via the reduction of unintentional point defects in the active region and waveguide, which reduces the non-radiative recombination by 2 orders of magnitude. A higher lasing threshold of 11 kW/cm2 is observed for AlGaN barriers, owing to the reduced localization of electrons and holes in the wells. It is shown that for electrically injected UVC laser diodes, AlGaN barriers are essential. © 2018 The Japan Society of Applied Physics

lectrically injected AlGaN-based UVC laser diodes are widely desired for application in spectroscopy, E non-line-of-sight communication, and biosensing.1,2) Thus far, optically pumped lasing in the UVC range has been demonstrated by several groups using both sapphire and AlN substrates.3–9) The achieved laser thresholds range from approximately 50 kW=cm2 to several MW=cm2. The dif- ferences in the laser threshold mainly arise from high point defect concentration due to unfavorable growth conditions, as well as high dislocation density in the case of growth on non- native substrates.10–12) Additionally, the design of the multi- ple quantum well (MQW) structure influences the perform- ance of the optically pumped lasers.13) Understanding and controlling all these parameters to achieve the lowest possible laser threshold and highest possible internal quantum effi- ciency (IQE) is crucial for achieving electrically injected = lasing. Among others, the performance of MQWs is impacted Fig. 1. STEM image of an Al0.55Ga0.45N Al0.65Ga0.35N MQW structure fi by (1) the Al content in the wells and barriers; (2) the with an EBL. All wells and barriers are well-de ned and of homogenous composition. thickness of the wells and barriers; (3) the number of QWs; and (4) the quality of the MQW structure, with regard to the homogeneity and interfacial roughness.14) However, despite many demonstrations of optically pumped lasing in the UV microscopy (STEM) via the Si-stacking procedure.18) STEM range, little is known about the impact of the MQW design imaging was performed using an FEI G2 60–300 kV Titan parameters on the laser threshold. operated at 200 kV with a 19.6 mrad convergence angle and In this study, 6 kW=cm2 threshold UVC lasing is demon- a 77 mrad collection semi-angle. A typical STEM image of strated via a reduction of the point defect concentration in the a sample with an Al content of 65% in the barriers is shown waveguide and in the MQW. The distinction between the in Fig. 1. All the wells and barriers are homogenous, and design of optically pumped and electrically driven laser struc- the interfaces are sharp and well-defined. For the depicted tures, as they must satisfy different requirements, is discussed. sample, the Al content in the barriers and waveguide is the Optically pumped lasers were grown via metalorganic same (65%). Band-structure potential fluctuations due to chemical vapor deposition on single-crystalline physical vapor inhomogeneous Al incorporation, as previously observed transport grown AlN substrates.3,15–17) First, a homoepitaxial in AlGaN MQWs grown on SiC and sapphire, are never AlN layer of approximately 200 nm thickness was grown. observed for structures grown on AlN substrates with proper Next, a 150-nm-thick AlGaN waveguide with an Al content surface supersaturation control.19–22) of 65% was deposited, followed by a 3× AlGaN=AlGaN After the growth, all samples were thinned and backside- MQW (2.5 nm=2.5 nm). The Al content in the wells for all polished to allow for reliable cleaving. The cleaved laser the investigated samples was 55% for a targeted emission of bars were mounted for photoluminescence (PL) and opti- approximately 265 nm, and the Al content in the barriers cally pumped laser measurements.23) An ArF (λ = 193 nm, was varied from 60 to 100%. The structure was finished by Ephoton = 6.4 eV) laser was used as an excitation source. The a 5-nm-thick AlN layer that was similar to the electron- beam intensity was reduced to a reasonable pumping power blocking layer (EBL) in electrically injected UV lasers and using several neutral-density filters and then focused onto served as a protection layer in the optically pumped lasers. the samples. The excitation power was measured using a Details regarding the growth can be found elsewhere.11) Coherent Powermax power meter. For optically pumped To assess the structural properties of the MQW structures, lasing, light was collected from the cleaved facets of the samples were prepared for scanning transmission electron structures using an optical fiber. Light dispersion and detec- 082101-1 © 2018 The Japan Society of Applied Physics Appl. Phys. Express 11, 082101 (2018) R. Kirste et al.

wavelength (nm) (a) 280 270 260 45k sample A (a) PL, ArF laser (b) 40k sample B sample A 35k 30k (b) 25k 20k 15k 10k (c) intensity (arb.units) 5k 300 K (arb.units) intensity 2 K 0 1101004.6 4.8 power density (kW/cm2) energy (eV)

Fig. 2. (a) UV light emission intensity of optically pumped lasers as a function of the excitation power density for a sample with AlN barriers, where the waveguide was grown under high (sample A) and low (sample B) Fig. 3. Schematics of conduction-band structures of three samples with supersaturation. It is found that a lower point defect concentration different barrier compositions: (a) barriers with a higher Al content than the (sample A) reduces the laser threshold by approximately 1 order of waveguide, (b) barriers with the same Al content as the waveguide, and magnitude. (b) Low-temperature (2 K) and room-temperature PL spectra of (c) barriers with a lower Al content than the waveguide. an AlGaN=AlN MQW (sample A). The room-temperature IQE of the MQW of 95% at a carrier concentration of 1 × 1018 cm−3 is estimated according to the temperature-dependent PL data. improvement over previously reported values.11) In contrast, sample B showed a significantly more drastic drop of the tion was performed using a 0.75-m single monochromator light intensity under the same illumination condition (not and a cooled charge-coupled device camera. shown). At 1 × 1018 cm−3 (50 kW=cm2), this sample exhib- The effect of the growth conditions on the internal quantum ited only 2=3 of the IQE of sample A, which agrees well with efficiency (IQE) has previously been discussed,11) showing previous results.11) Both the IQE and laser threshold meas- that an increase in supersaturation via higher V=III ratios urements confirm that point defects have a significant impact led to higher IQEs for a given carrier concentration owing on the performance of MQWs and therefore UV light- to a decrease in the non-radiative recombination. This dem- emitting diodes and lasers. It is noted that the impact of the onstrated that in addition to dislocations, point defects were point defects on the laser threshold is more pronounced than another major factor that strongly influenced the optical the impact on the IQE. It is argued that this is because point quality of AlGaN-based UV lasers. Therefore, to reduce the defects not only impact the free carriers and generation laser threshold of optically pumped UV lasers, further reduc- of light via radiative recombination but also cause optical tion of point defects in the MQW and the AlGaN waveguide loss=modal loss via absorption of the laser light waves. This was pursued. Figure 2(a) compares the laser threshold of has recently been investigated experimentally and theoret- an optically pumped laser where the waveguide and MQW ically for UV laser devices.26) Finally, it should be pointed were grown under V=III ratios of 1,300 (sample A) and 400 out that a reduction of the point defect concentration can (sample B) by varying the ammonia flow rate. This represents be achieved using tools besides the V=III ratio, including high and low supersaturation, respectively. Increasing the supersaturation (temperature, carrier gas, pressure, etc.) or supersaturation by increasing the V=III ratio is expected to external measures such as Fermi-level control.24,27) reduce the point defect concentration and the non-radiative On the basis of the low-point defect density sample A, recombination in sample A.11,24,25) a series of MQWs with varying Al content in the barriers was Both laser structures emitted light at 267 nm. From grown to investigate the influence of the barrier composition Fig. 2(a), it is found that both samples show the typical in the MQW on the carrier recombination and emission laser turn-on behavior that is expected for optically pumped properties. Figure 3 shows schematics of the conduction lasing.3) In addition, a reduction in the full width at half maxi- band structures of samples with AlN barriers (a), as well as mum (FWHM) of the emission peak to approximately 1 nm AlGaN barriers with Al contents of 65% (b) and 60% (c). and 100% transverse-electric polarization of the emitted light Because the waveguide composition was 65% Al, the first was observed above the laser threshold. However, it is found sample had barriers with a higher Al content than the wave- that the sample with the lower point defect concentration guide, the barriers for the second sample had the same Al (sample A) has a threshold of 6 kW=cm2, which is a reduc- content as the waveguide, and the third sample had a lower tion of 1 order of magnitude compared with sample B Al content in the barriers than that in the waveguide. (60 kW=cm2). Figure 2(b) shows room-temperature and In room-temperature PL measurements, all samples low-temperature (2 K) PL spectra for sample A. When the exhibited a strong emission around 265 nm. The FWHM of temperature is increased from 2 K to room temperature, the the PL emission peaks at room temperature was approx- intensity of the MQW emission is only slightly reduced. imately 12 nm. When luminescence was collected from the According to the observed intensity reduction, the IQE of the cleaved facets, optically pumped lasing was observed for all MQW is determined to be 95% at an estimated carrier density samples. However, different excitation power densities were of 1 × 1018 cm−3 (50 kW=cm2). This result was confirmed needed to achieve lasing, depending on the Al content in the using power-dependent IQE measurements and is a major barriers. Figure 4 shows the emission intensity of the MQW 082101-2 © 2018 The Japan Society of Applied Physics Appl. Phys. Express 11, 082101 (2018) R. Kirste et al.

solid column - electron confined charge 21 60k barrier AlN ) 10 patterned column - hole confined charge barrier 65% -3 20 AlN AlN AlN 50k barrier 60% 10 65% 65% 65% 60% 60% 60% 1019 40k 1018 30k 1017

20k 16 intensity (cts) 10 10k 1015 1014

0 (cm density carrier confined QW1 QW2 QW3 5 10152025303540 power density (kW/cm2) quantum well

Fig. 4. Emission intensity of three different MQW structures as a function Fig. 5. Electron and hole density (solid bars and striped bars, respectively) of the excitation power density. The laser threshold increases with the in the three AlGaN QWs with different barriers. For the AlN barriers (black decreasing barrier height. bars), electrons are found only in the first QW, while the use of AlGaN barriers (green and blue bars) leads to sufficient electron and hole injection into all three QWs. The laser structure was simulated using the ATLAS framework (SILVACO). peaks as a function of the excitation power density for the three different barrier configurations. All three samples show clear lasing threshold behavior with a well-defined nonlinear 65% (green), and AlGaN barriers with an Al content of 60% intensity increase above the threshold. The sample with the (blue). The applied voltage in all three cases was kept con- lowest threshold is the same as sample A in Fig. 2. Com- stant. In the structure with the AlN barriers, the electrons are paring the threshold of the different lasers reveals a clear injected primarily into QW1 (QW adjacent to the n-doped trend towards a lower threshold for higher barriers. Accord- waveguide), and the holes are injected into QW3 (QW ingly, the highest threshold is observed for the sample with adjacent to the p-doped waveguide). Insufficient electron an Al content of 60% in the barriers (32 kW=cm2), which is injection is predicted for the simulated device with the AlN approximately five times higher than the threshold of the barriers. While the electron concentration in QW1 (black sample with AlN barriers (6 kW=cm2). The structure with solid bar) is on the order of 1020 cm−3, in QW2, the concen- 65% Al in the AlGaN barriers has a threshold of 11 kW=cm2. tration has already dropped below 1015 cm−3. The predictions Because the wells in the investigated MQWs are relatively for the hole injection are slightly better; however, a drop of thin (2.5 nm), some leakage of the electron and hole wave 3 orders of magnitude is observed when QW3 and QW1 are function into the barriers can be expected.28) This leakage compared. This drop in the carrier concentration is due is reduced if the Al content in the barriers is increased. to the significant conduction-band and valence-band offset Therefore, the origin of the difference in the threshold is between AlN and AlGaN with 65% Al.29) Thus, it becomes likely the better confinement of carriers in the wells sur- obvious that because the conduction- and valence-band rounded by higher barriers, leading to lower leakage into the offsets in the AlGaN heterojunctions are split roughly 2=3 barriers. Other factors that may contribute to the improved and 1=3 of the bandgap difference,29,30) AlN acts as a very performance include the material gain, oscillator strength, effective EBL. Consequently, to increase the carrier injection and carrier lifetime. Lastly, it should be mentioned that even into MQWs for UV lasers, AlGaN barriers are mandatory. As for the lasers with the lowest barriers, the threshold is signi- shown in Fig. 5, neither hole nor electron injection should ficantly better than that of the structures with a higher point be a significant challenge in AlGaN=AlGaN-based MQWs if defect concentration (sample B in Fig. 2), which highlights the composition contrast between the wells and barriers is the need for clean epitaxial layers. This demonstrates that to approximately 10%, which results in approximately 200- and achieve electrically injected UV lasers, control of the epitaxy 100-meV barriers for the electrons and holes, respectively. and reduction of defects are required before the optimization A lower contrast results in poorer carrier confinement but a of the device structure. more uniform carrier distribution across all three QWs, as The experimental data clearly indicate that AlN barriers are shown in Fig. 5. advantageous for improving the carrier confinement and In summary, we investigated the influence of point defects reducing the lasing threshold for optically pumped lasers. and the MQW design on the optical and electrical properties However, optically pumped lasing does not consider carrier of UV laser structures. A low optically pumped laser threshold injection into the MQW under conditions important for elec- of 6 kW=cm2 was found for AlGaN-based MQWs with AlN trically injected UV laser diodes. This is demonstrated by barriers. A 1 order of magnitude decrease in the lasing data obtained from the simulation (ATLAS) of complete laser threshold compared with the literature data was achieved by structures (700 nm n-cladding=60 nm n-waveguide=MQW= controlling the point defects in the MQW and in the under- 60 nm p-waveguide=60 nm p-cladding=p-GaN–the Al-con- lying waveguide. The V=III ratio is only one pathway for tent in the cladding layers is 70% and in the waveguide is reducing the point defect-concentration in devices. Others 65%). As an example, Fig. 5 shows the number of confined include supersaturation (temperature, carrier gas, pressure, electrons (solid columns) and holes (striped columns) in the etc.) and external control, such as Fermi-level control.24,27) active region of a laser diode with three wells separated by For barriers with a lower Al-content, a slightly increased laser AlN barriers (black), AlGaN barriers with an Al content of threshold was observed. This observation is explained by the 082101-3 © 2018 The Japan Society of Applied Physics Appl. Phys. Express 11, 082101 (2018) R. Kirste et al. better carrier confinement under optical pumping for the 9) D. Li, K. Jiang, X. Sun, and C. Guo, Adv. Opt. Photonics 10, 43 (2018). 326 MQWs with high barriers. Maximization of the carrier con- 10) H. Amano, J. Phys.: Conf. Ser. , 012002 (2011). 11) Z. Bryan, I. Bryan, J. Xie, S. Mita, Z. Sitar, and R. Collazo, Appl. Phys. finement is consistent with the design of optically pumped Lett. 106, 142107 (2015). laser structures. However, for electrically driven lasers, carrier 12) M. Shatalov, W. Sun, A. Lunev, X. Hu, A. Dobrinsky, Y. Bilenko, J. Yang, injection into MQWs with AlN barriers is insufficient owing M. Shur, R. Gaska, C. Moe, G. Garrett, and M. Wraback, Appl. Phys. fi ff Express 5, 082101 (2012). to the signi cant band o set between the AlN and AlGaN 13) W. W. Chow, M. Kneissl, J. E. Northrup, and N. M. Johnson, Appl. Phys. with 55% Al, and it is necessary to use lower barriers to ensure Lett. 90, 101116 (2007). sufficient carrier injection into all QWs. Lowering of the 14) S. Nakamura, S. Pearton, and G. Fasol, The Blue Laser Diode: The barriers increases the threshold, as discussed previously. Complete Story (Springer, Berlin, 2013). ff 15) R. Dalmau, B. Moody, R. Schlesser, S. Mita, J. Xie, M. Feneberg, B. However, Figs. 4 and 5 show that the e ect is marginal. Neuschl, K. Thonke, R. Collazo, A. Rice, J. Tweedie, and Z. Sitar, Acknowledgments The authors acknowledge partial financial support J. Electrochem. Soc. 158, H530 (2011). from ARL (W56KGU-15-C-0052, W56KGU-14-C-0046), NSF (ECCS-1508854, 16) A. Rice, R. Collazo, J. Tweedie, R. Dalmau, S. Mita, J. Xie, and Z. Sitar, ECCS-1610992, DMR-1508191, ECCS-1653383), and ARO (W911NF-15-2- J. Appl. Phys. 108, 043510 (2010). 0068, W911NF-16-C-0101). Part of this work was performed at the Analytical 17) D. Ehrentraut and Z. Sitar, MRS Bull. 34, 259 (2009). Instrumentation Facility (AIF) of North Carolina State University, which is 18) J. H. Dycus and J. M. Lebeau, J. Microsc. 268, 225 (2017). supported by the State of North Carolina and the National Science Foundation 19) E. F. Pecora, W. Zhang, A. Y. Nikiforov, L. Zhou, D. J. Smith, J. Yin, R. (award number ECCS-1542015). The AIF is a member of the North Carolina Paiella, L. D. Negro, and T. D. Moustakas, Appl. Phys. Lett. 100, 061111 Research Triangle Nanotechnology Network (RTNN), which is a site in the (2012). National Nanotechnology Coordinated Infrastructure (NNCI). 20) Y. Iwata, T. Oto, D. Gachet, R. G. Banal, M. Funato, and Y. Kawakami, J. Appl. Phys. 117, 115702 (2015). 21) I. Bryan, Z. Bryan, S. Mita, A. Rice, J. Tweedie, R. Collazo, and Z. Sitar, J. Cryst. Growth 438, 81 (2016). 1) M. Kneissl, in III–Nitride Ultraviolet Emitters: Technology and 22) I. Bryan, Z. Bryan, S. Mita, A. Rice, L. Hussey, C. Shelton, J. Tweedie, Applications, ed. M. Kneissl and J. Rass (Springer, Cham, 2016) Chap. 1. J.-P. Maria, R. Collazo, and Z. Sitar, J. Cryst. Growth 451, 65 (2016). 2) R. Kirste, N. Rohrbaugh, I. Bryan, Z. Bryan, R. Collazo, and A. Ivanisevic, 23) W. Guo, Z. Bryan, J. Q. Xie, R. Kirste, S. Mita, I. Bryan, L. Hussey, M. Annu. Rev. Anal. Chem. 8, 149 (2015). Bobea, B. Haidet, M. Gerhold, R. Collazo, and Z. Sitar, J. Appl. Phys. 115, 3) J. Xie, S. Mita, Z. Bryan, W. Guo, L. Hussey, B. Moody, R. Schlesser, R. 103108 (2014). Kirste, M. Gerhold, R. Collazo, and Z. Sitar, Appl. Phys. Lett. 102, 171102 24) F. Kaess, S. Mita, J. Xie, P. Reddy, A. Klump, L. H. Hernandez- (2013). Balderrama, S. Washiyama, A. Franke, R. Kirste, A. Hoffmann, R. Collazo, 4) T. Wunderer, C. L. Chua, J. E. Northrup, Z. Yang, N. M. Johnson, M. and Z. Sitar, J. Appl. Phys. 120, 105701 (2016). Kneissl, G. A. Garrett, H. Shen, M. Wraback, B. Moody, H. S. Craft, R. 25) P. Reddy, F. Kaess, J. Tweedie, R. Kirste, S. Mita, R. Collazo, and Z. Sitar, Schlesser, R. F. Dalmau, and Z. Sitar, Phys. Status Solidi C 9, 822 (2012). Appl. Phys. Lett. 111, 152101 (2017). 5) F. Asif, M. Lachab, A. Coleman, I. Ahmad, B. Zhang, V. Adivarahan, and 26) M. Martens, C. Kuhn, T. Simoneit, S. Hagedorn, A. Knauer, T. Wernicke, A. Khan, J. Vac. Sci. Technol. B 32, 061204 (2014). M. Weyers, and M. Kneissl, Appl. Phys. Lett. 110, 081103 (2017). 6) X.-H. Li, T. Detchprohm, T.-T. Kao, M. M. Satter, S.-C. Shen, P. D. Yoder, 27) S. Mita, R. Collazo, A. Rice, R. F. Dalmau, and Z. Sitar, J. Appl. Phys. 104, R. D. Dupuis, S. Wang, Y. O. Wei, H. Xie, A. M. Fischer, F. A. Ponce, T. 013521 (2008). Wernicke, C. Reich, M. Martens, and M. Kneissl, Appl. Phys. Lett. 105, 28) K. Kojima, K. Furusawa, Y. Yamazaki, H. Miyake, K. Hiramatsu, and S. F. 141106 (2014). Chichibu, Appl. Phys. Express 10, 015802 (2017). 7) J. Zhang, H. Zhao, and N. Tansu, Appl. Phys. Lett. 98, 171111 (2011). 29) P. Reddy, I. Bryan, Z. Bryan, J. Tweedie, S. Washiyama, R. Kirste, S. Mita, 8) Y.-S. Liu, Z. Lochner, T.-T. Kao, M. M. Satter, X.-H. Li, J.-H. Ryou, S.-C. R. Collazo, and Z. Sitar, Appl. Phys. Lett. 107, 091603 (2015). Shen, P. D. Yoder, T. Detchprohm, R. D. Dupuis, Y. Wei, H. Xie, A. 30) Z. Sitar, M. J. Paisley, B. Yan, R. F. Davis, J. Ruan, and J. W. Choyke, Thin Fischer, and F. Ponce, Phys. Status Solidi C 11, 258 (2014). Solid Films 200, 311 (1991).

082101-4 © 2018 The Japan Society of Applied Physics